The Mayo Alzheimer's Disease Patient Registry (ADPR) was initiated in 1986 in response to an RFA from the National Institute on Aging to establish registries for the study of the incidence, prevalence, and risk factors for AIzheimer's Disease (AD) and to refine diagnostic criteria for dementia. Mayo responded by proposing a registry comprised of two sets of projects: One set of projects was designed to use the records-linkage system of the Rochester Epidemiology Project to study incidence, prevalence, and risk factors in the population of Rochester, Minnesota. The other set of projects was designed to establish a prospective, longitudinal cohort of normal and cognitively impaired subjects and to refine the diagnostic criteria for cognitive impairment. This cohort continues to be followed with some subjects having up to15 years of longitudinal data. From this work has come a wealth of research on normal aging, mild cognitive impairment (MCl), and AD, including studies on the normative neuropsychology, neuroimaging, genetics, and neuropathology of this cohort. In this renewal application, the same themes concerning the incidence, prevalence, risk factors, and diagnostic criteria for cognitive impairment and dementia will be pursued; however, there will be a major innovation in methodology. On the advice of our External Advisory Committee, we prepare to use the resources of the Rochester Epidemiology Project to establish a population-based cohort of subjects aged 70-89 years. We plan to recruit a stratified random sample of 2,300 subjects from OImsted County, MN, to study them via telephone contact and direct examination at baseline, and to follow them prospectively by repeated examination annually. Subjects who refuse participation and those who have incomplete follow-up will be studied passively through the records-linkage system. MCl, dementia, and AD will be defined using existing sets of diagnostic criteria. We propose to study the prevalence of MCl and other forms of intermediate cognitive impairment, the incidence of MCI, and to study risk factors or predictors for cognitive impairment. As the field of aging and dementia moves toward the earlier identification of cognitive impairment with the goal of developing prevention strategies, questions regarding the incidence, prevalence, and risk factors for MCl become increasingly important. This population-based cohort will be a valuable resource for additional research on epidemiology, clinical characteristics, neuropsychological features, neuroimaging measures, and neuropathological substrate of prodromal dementia.